My general practice when sharing Jupyter notebooks has been to download them as HTML and hide the code blocks - this works well because it preserves the interactive aspects of the Bokeh charts.
After I recently updated Bokeh, none of the charts I am creating are visible when I download a Jupyter notebook as HTML. I am running Bokeh 0.12.1, Python 3.5, and iPython 5.1.0. Any solutions to this issue?
Update: regressing to Bokeh 0.11.1 and iPython 5.0.0 resolves this issue, and also resolves other ongoing issues around needing to run output_notebook() multiple times.
In order for Bokeh to display anything at all, the BokehJS client library must be loaded. BokehsJS is loaded by the JavaScript code that is executed in the notebook output cell for output_notebook. If you hide the cell containing output_notebook is hidden, the code to load BokehJS is never run, BokehJS is not loaded, and no Bokeh plots will display.
Long story short: you can't hide the cell that calls output_notebook
See this issue for further discussion, as well as a possible alternative to hiding using "notes" cells (it's mentioned there but I don't know anything more about it)
Related
I am using nbinteract to develop an interactive web page from a Jupyter notebook. I finally got to the end and published the first version of it but it does not appear the python libraries loaded properly (see image below). This appears to be the problem even in the original nbinteract tutorial. Any ideas on what might be the problem here?
Thank you
It’s possible to export a notebook including widgets to html using nbconvert (https://github.com/jupyter/nbconvert ). You need to make sure to save the notebook using the classic Jupyter Notebook (not JupyterLab) with the “Widgets -> Save Notebook Widget State” menu option.
Unfortunately, it’s not possible to preserve the behavior of the callback functions this way because these functions are defined using Python, and there’s no Python kernel available in the standalone HTML file.
To host interactive figures outside of the notebook , writing a Dash app is always good (https://dash.plot.ly/ ). or If you want to stay in notebook/widget field you could use https://github.com/QuantStack/voila.
I'm graphing some stuff with matplotlib and I need to be able to zoom in on the graph. But VSCode just renders it in its Python Interactive window that I can't do anything with. Is there a way to open graphs in a their own window that allows for scaling and things?
I'm a developer on this extension. We don't currently support popping out graphs. If you want to vote up the issue, we have an issue filed for this support here:
https://github.com/Microsoft/vscode-python/issues/4976
Until this support is added there is a bit of a work around. You can manually use the %matplotlib ipython magic to update to non-inline rendering. Then when you show plots they will be shown in a popup window. Like so:
I can't tell you exactly what command to use as it might vary based on what you have installed in your environment. But %matplotlib auto or %matplotlib qt would be good places to try starting with.
Plotly says it is a web library. Matplotlib does support Wxpython but comparatively, it is too low level. Is there any way to integrate Plotly inside a Wxpython GUI or are there any alternative for it?
You might be able to use Plotly's ability to export plots to images to get the plots into wxPython:
https://plot.ly/python/offline/#static-image-export
If you need the live plotting though, then you will need to use Webview:
https://wxpython.org/Phoenix/docs/html/wx.html2.WebView.html
I also see that Plotly can now be used offline, so if you used that plus Webview, it should work.
Wxpython webview doesn't support 3-d graphs from plotly.
For 3-d graphs, I used cefpython. It embeds a full chromium browser inside python application. wxpython example can be found here: https://github.com/cztomczak/cefpython/blob/master/examples/wxpython.py.
I used plotly offline, and read the file inside cefpython.
I used this answer to build a requirements.txt file for my Jupyter notebook. I then put both the ipynb and requirements.txt file in a Git repo and published it to Binder. As I understand, this should give me an interactive version of my notebook which I can share with people for them to play around with.
The published Binder can be found here.
Does anyone know why the interactive bit is not showing? Specifically, the sliders.
You need to enable the extension that allows the sliders, namely ipywidgets. There is an example of using ipywidgets presently here among the Binder example template repos. (If that repo gets altered, I was talking about this specific commit.)
Right now the extension gets enabled separately for JupyterLab vs the classic interface. If you just want to have your launches default to the classic interface, you can leave out the JupyterLab-related enabling.
Usually I do plotting inside of IPython Notebook with pylab mode.
Whenever I use Bokeh, I like to enable output_notebook() to show my plot inside of the IPython notebook.
Most annoying part is that Bokeh enable wheel_zoom by default which cause unintended zoom in IPython notebook.
I know I can avoid this by passing comma separated tools string what I want to include into bokeh.plotting.figure. but with this solution, I should list up the other tools but wheel_zoom.
Is there any way to exclude wheel_zoom only? or Can I disable wheel_zoom in global setting or something like that?
There is an open PR to improve this, it will be in the the 0.11 release.